- Email: [email protected]
Quality assessment of marketed chamomile tea products by a validated HPTLC method combined with multivariate analysis
Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Contents lists available at ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba
Quality assessment of marketed chamomile tea products by a validated HPTLC method combined with multivariate analysis d ´ Etil Guzelmeric a , Petar Ristivojevic´ b , Irena Vovk c , Duˇsanka Milojkovic-Opsenica , a,∗ Erdem Yesilada a
Yeditepe University, Faculty of Pharmacy, Department of Pharmacognosy and Phytotherapy, Kayisdagi Cad., Atasehir, 34755 Istanbul, Turkey Innovation Centre of Faculty of Chemistry Ltd., Studenski trg 12-16, 11000 Belgrade, Serbia c Department of Food Chemistry, National Institute of Chemistry, Hajdrihova 19, SI-1000 Ljubljana, Slovenia d Faculty of Chemistry, University of Belgrade, Studentski trg 12-16, 11158 Belgrade, Serbia b
a r t i c l e
i n f o
Article history: Received 15 June 2016 Received in revised form 19 September 2016 Accepted 24 September 2016 Available online 26 September 2016 Keywords: Matricaria recutita L. High performance thin-layer chromatography (HPTLC) Apigenin 7-O-glucoside Quality control Chamomile tea products Multivariate analysis
a b s t r a c t Chamomile tea composed of dried flower heads of Matricaria recutita L. (Asteraceae) is one of the most popular single ingredient herbal teas. Tea industries, spice shops or public bazaars are mostly supplied chamomile as a raw material via cultivation or through nature-picking. However, one of the drawbacks of nature-picking is adulteration. This could be either due to false authentication of the plant materials by ingenuous pickers or intentional/unintentional substitution with other flowers resembling to chamomile in appearance during harvesting. Therefore, quality control of raw chamomile materials before marketing should be carefully considered not only by quantification of apigenin 7-O-glucoside (active marker) but also by fingerprinting of chemical composition. This work presents both quantification of apigenin 7-Oglucoside and chemical fingerprinting of commercial chamomile tea products obtained from different food stores and spice shops by a validated HPTLC method. In addition, HPTLC profiles of investigated chamomile tea samples were compared with HPLC method stated in the European Pharmacopoeia and it was found that HPTLC method was superior to HPLC method in the field of adulteration confirmation. Therefore, fingerprint profiles performed on the silica gel 60 NH2 F254 s HPTLC plates combined with pattern recognition techniques of these marketed products were comparatively evaluated with wild and cultivar chamomile samples and also chamomile-like species from Asteraceae. Consequently, not chamomile tea bags but crude flowers sold on market were found to be adulterated with other plant materials. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Brewed tea of chamomile flowers (Matricaria recutita L., Asteraceae) has been consumed for centuries either for its delight or healing benefits such as alleviation of the symptoms of gastrointestinal complaints, inflammatory and nervous problems. Apart from its traditional use, recent investigations on health promoting efficacy of chamomile tea have also revealed its antioxidant, antinociceptive, anticancer etc. activities mostly associated with one of its major components, apigenin 7-O-glucoside (A7G) [1,2]. Today, daily more than one million cups of chamomile tea have been consumed throughout the world [3]. Tea industries, food stores, spice shops or public bazaars are mostly supplied
∗ Corresponding author. E-mail address: [email protected] (E. Yesilada). http://dx.doi.org/10.1016/j.jpba.2016.09.030 0731-7085/© 2016 Elsevier B.V. All rights reserved.
chamomile flowers as a raw material in two ways: cultivation or through nature-picking. Then, they introduce them into market both crude (bulk or wrapped) and processed (tea bags). Actually, retailers commonly prefer nature-picking materials due to its lower cost. However, one of the drawbacks of such materials is adulteration due to false authentication of the plant materials by ingenuous pickers or intentional/unintentional substitution with other flowers resembling to chamomile in appearance during harvesting [4]. On the other hand, recent ethnobotanical surveys carried out in different localities of Turkey revealed that several other species from Asteraceae with similar appearance are also named “papatya” [Turkish name of chamomile] i.e. Anthemis tinctoria L., Anthemis cretica L., Anthemis cotula L., Anthemis austriaca L., Anthemis ceolopoda L., Anthemis wiedemanniana L., Bellis perennis L., Chrysanthemum coronarium L. [5–10]. Even more important point, severe problems may be emerged by the use of such false chamomile materials i.e. Senecio sp., which is reported to induce
36
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Table 1 Analyzed chamomile samples in this study. Number of the samples
Samples
Sample Codes
Detailed Information
5 11 14
Tea bags Bulk or packaged crude flowers sold as ‘chamomile’ Wild chamomiles and chamomile-like flowers belonging to 4 different genera, Anthemis L., Bellis L., Tanacetum L. and Chrysanthemum L. Bona, Bodegold and Zloyt Lan varieties of chamomile cultivars and one cultivar from local population (Yalova) having 3 different sowing dates (early November, early December and late December)
M1-5 A1-11 –
purchased from food stores (Istanbul, Turkey) purchased from bazaar (Istanbul, Turkey) gathered from nature in different localities of Turkey by local people assuming as chamomile.
–
obtained from Ataturk Central Horticultural Research Institute (Yalova, Turkey)
12
severe hepatic veno-occlusive disease and even death [11]. Moreover, loss in the expected efficacy might be the other negative impact due to low yield or absence of active ingredient in the false materials. The widespread usage and well-known popularity of chamomile flowers have raisen it to be documented in the pharmacopoeias of 26 countries including European Pharmacopoeia (Ph. Eur.) [12]. In the Ph. Eur., A7G is stated as an active marker of chamomile flowers and an high-performance liquid chromatographic (HPLC) method is described for estimation of its total concentration [13]. High performance thin-layer chromatography (HPTLC) as an analytical technique provides convenience for both qualitative and quantitative conclusions. Nowadays, HPTLC fingerprints in combination with pattern recognition methods has become popular either to provide useful information on complex matrices or to support analytical results [14–16]. The aim of the present study was to: 1.) reveal chemical fingerprint profiles of the marketed chamomile tea products and quantify A7G in these samples using previously reported validated HPTLC method [17]; 2.) compare the selectivity of the HPTLC method and HPLC method stated in the Ph. Eur. in discrimination of the genuine specimen; 3.) analyze and evaluate HPTLC fingerprint profiles of marketed chamomile tea products with those of wild and cultivar chamomile samples, as well as other species having chamomile-like flowers (Anthemis spp., Bellis spp., Tanacetum sp., Chrysanthemum sp.) from the family Asteraceae using pattern recognition techniques.
2. Experimental
Table 2 Commercial chamomile tea products and chamomile cultivars: yields (%) after lyophilization and A7G contents.
Tea products
Cultivars
Product
Yield (%)
A7G (mg/g)x
M1 M2 M3 M4 M5 A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 Bodegolda Bodegoldb Bodegoldc Bonaa Bonab Bonac Zloyt Lana Zloyt Lanb Zloyt Lanc Yalovaa Yalovab Yalovac
17.49 20.24 24.10 19.82 28.37 24.07 28.63 19.14 20.35 14.98 15.97 17.64 21.62 20.04 12.70 13.84 34.05 32.47 29.32 30.07 31.09 29.40 29.83 32.25 31.8 42.05 39.19 36.05
0.49 ± 0.01 0.49 ± 0.03 0.80 ± 0.02 0.54 ± 0.01 0.43 ± 0.01 0.30 ± 0.02 0.51 ± 0.01
ndt: A7G not detected; < LOQ: under the limit of quantitation (5 ng/band). a,b,c Sowing dates starting by early November (a), early December (b) and late December (c), respectively. x Mean ± SD (n = 3).
2.1. Chemicals HPLC grade acetonitrile was obtained from J. T. Baker (Deventer, the Netherlands). The other solvents were of analytical grade. Ethanol absolute, ethyl acetate and formic acid were purchased from Sigma-Aldrich (Steinheim, Germany); acetic acid was from Riedel-de Haen (Seelze, Germany), o-phosphoric acid was from Merck (Darmstadt, Germany); methanol and dichloromethane were from Analar Normapur (Muarrie, Australia). 2-aminoethyl diphenylborinate and polyethylene glycol 400 were from Fluka (Steinheim, Germany) and Merck (Hohenbrunn, Germany), respectively. The ultrapure water was obtained from Millipore, Simplicity UV (Darmstadt, Germany). Sodium hydroxide, magnesium chloride-6-hydrate as well as standards of A7G and 5,7-dihydroxy-4-methylcoumarin were purchased from Sigma-Aldrich (Steinheim, Germany).
2.2. Plant materials The analyzed samples were chamomile tea bags, crude chamomile flowers, wild and cultivar chamomiles as well as several chamomile-like flowers from Asteraceae (Table 1).
The materials stored in a refrigerator at −25◦ C were ground to powder in a mechanic grinder before extraction. In addition, six tea bags from each different tea brands (M1-5) being in a powdered form were randomly selected due to variable weight (1.3–2.0 g) and mixed in a beaker for homogenous sampling.
2.3. Preparation of standard solutions 2.3.1. Standard solutions for HPTLC analysis A7G stock solution (250 g/mL) was prepared in methanol and further diluted with the same solvent to prepare the working solutions 2.5, 5 and 10 g/mL.
2.3.2. Standard solutions for HPLC analysis A7G standard solution (12.5 g/mL) and standard solution mixture containing both A7G and 5,7-dihydroxy-4-methylcoumarin (10 g/mL) were prepared according to the procedure described in the Ph. Eur. [13].
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
2.4. Preparation of sample test solutions 2.4.1. Extraction of wild and cultivar chamomiles, chamomile-like flowers and chamomile tea products for HPTLC analysis All samples were brewed to prepare 2% infusion following our previously described procedure [17]. Next, they were filtered, cooled and finally freeze-dried. The yields of the lyophilized commercial and cultivar samples are listed in Table 2. The yields of other samples are stated in the previous report [17]. Then, 10 mg of each wild chamomile and chamomile-like sample and 20 mg of each chamomile tea product was accurately weighed and extracted with 10 mL of methanol in an ultrasonic bath for 15 min. Suspended particles were removed by filtration through a 0.45 m RC-membrane filter (Sartorius stedim biotech). Further, the sample test solutions of chamomile cultivars were diluted 2 times.
37
HPLC system suitability confirmed by the resolution (Rs ) between the standards A7G and 5,7-dihydroxy-4-methylcoumarin which was found as 7.54 ± 0.03 (n = 3). The result was found to be suitable for further analyses according to the Ph. Eur. The retention time (tR ) of the A7G standard (tR = 6.35 ± 0.05; n = 3), presented in Fig. 3 were compared with the tR obtained for the reference chamomile sample solution to evaluate the identity of the A7G. 2.7. Statistical analysis The quantitative results of HPTLC analyses of each sample test solution applied in triplicate on the same plate were expressed as mean ± standard deviation (SD). These treatments were carried out by using SPSS Data Editor (version 20.0). 2.8. Image processing and multivariate analysis
2.4.2. Extraction of reference chamomile and chamomile tea products for HPLC analysis One gram of the powdered chamomile flowers were accurately weighed and extracted with 100 mL EtOH (96%), and then heated under a reflux condenser on a water-bath for 15 min and cooled and filtered with a filter paper. 10 mL of freshly prepared diluted sodium hydroxide solution was added to the filtrate and the mixture was heated under a reflux condenser on a water-bath for 1 h and cooled. It was diluted to 125 mL with EtOH (96%). To 25 mL of the solution, 0.25 g of citric acid was added and shook for 5 min. First, it was filtered with a filter paper and then filtered through a 0.45 m RCmembrane filter. 5 mL of this solution diluted to 10 mL with the initial mixture of the mobile phase [13]. 2.5. HPTLC method All samples were analyzed by previously validated HPTLC method [17]. HPTLC glass plates (20 cm × 10 cm) precoated with silica gel 60 NH2 F254 s (Merck), Linomat V automatic sample spotter (Camag, Muttenz, Switzerland) equipped with 100 L Hamilton syringe (Bonaduz, Switzerland), Automatic Developing Chamber 2 (ADC2, Camag) and TLC Scanner 3 (Camag) were used during analyses. After development with ethyl acetate-formic acid-acetic acid-water (30:1.5:1.5:3, v/v/v/v), the plates were heated at 100◦ C on the Camag TLC plate heater for 3 min and dipped into Natural Products (NP) and Polyethylene glycol (PEG) 400 solutions prepared as stated by Reich and Schibli [18]. After, the plates were documented by the Camag TLC visualizer at 366 nm. All the instruments were operated by winCATS program (version 1.4.8, Camag). The identity of the A7G in all samples was evaluated by comparison of the retention factors (RF ) of the sample zone and the zone of A7G standard, 0.38 ± 0.01, (Figs. 1 and 2). For the quantitative analysis, 5, 10, 15, 20, 25 and 50 ng/band A7G standard solution was applied on the plate for the calibration curve and A7G content was evaluated in all samples through peak area via quadratic regression (R2 = 0.9995). The application volume of each sample test solutions was 2 L. 2.6. HPLC method HPLC analyses were performed as described in the Ph. Eur. monograph for chamomile [13] and carried out by using the following equipments; Agilent 1100 HPLC system (Technologies, Santa Clara, California, USA) coupled with ChemStation 10.01 software, a model G1379A vacuum degasser, a model G1311A quaternary pump, a model G1313A auto-sampler, a model G1316A thermostated column compartment, and a model G1315 B diode array detector. Separations were achieved on an Agilent Zorbax Eclipse Plus C18 ODS column (250 mm × 4.6 mm, 5 m particle size).
Images of the HPTLC chromatograms were transformed to data set which contains 42 samples per 459 pixels for red, green and blue channels. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) as pattern recognition techniques were performed by using PLS Toolbox, v.6.2.1, for MATLAB 7.12.0 (R2011a) [19]. PCA was performed by using a singular value decomposition algorithm and a 0.95 confidence level for Q and T2 Hostelling limits for outliers. HCA was obtained by using the ward method to calculate cluster distances and applying Euclidean distance as a measure of distance between the samples. All data obtained from HPTLC chromatograms were preprocessed by using mean centered, standard normal variety (SNV) prior to any multivariate analysis. Correlation optimised warping (COW) was applied to correct the inter- and intra-plate peak shift due to variations in experimental conditions, analyst error and instrumental instability [20]. 3. Results and discussion 3.1. Quality assessment of marketed chamomile tea products For centuries, chamomile tea has been consumed due to either its pleasant taste or medicinal purposes. In the light of the foregoing, all samples were prepared by brewing at 100◦ C. Recently, Harbourne et al. investigated the A7G content of chamomile teas at different preparation temparatures through steeping in 100 mL of distilled water at 57, 70, 80, 90 and 100◦ C. They reported that the A7G concentration incrementally increased between 57 and 90◦ C and its content became stable from 90◦ C to 100◦ C [21]. Their optimised extraction procedure yielding the highest A7G content was found to be in accordance with the results of the present study. Moreover, the preferred extraction in this study was also compared with described sample preparation for HPLC analysis stated in the Ph. Eur. [13]. It should be noted that in the Ph. Eur. instead of free A7G content, total A7G amount (including monoor diacetylated derivatives of A7G) is estimated in hydroalcoholic chamomile extract. Therefore, after extraction of powdered chamomile flowers with 96% ethanol, the extract was subjected to HPTLC analysis both directly and following the ester hydrolysis by addition sodium hydroxide reagent to the extract and heating under a reflux condenser on a water-bath for 1 h. According to the HPTLC chromatogram in Fig. 4a (Track 1), many components disappeared after addition of sodium hydroxide possibly due to ester hydrolysis, Fig. 4a (Track 2). Moreover, the zones around RF = 0.60 and RF = 0.70 having characteristics color similar to that of A7G in Fig. 4a (Track 1) were not seen after hydrolysis in Figs. 4b and 4c; these zones could be mono- or diacetylated A7G derivatives. Fonseca and Tavares and Haghi et al. also pointed out that following ester hydrolysis, the peak area of A7G expanded [22,23]. On the
38
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Table 3 HPTLC densitograms of commercial chamomile tea products and chamomile cultivars.
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
39
Table 3 (Continued)
a,b,c
Sowing dates starting by early November (a), early December (b) and late December (c), respectively.
Fig. 1. HPTLC NH2 F254 s chromatogram of A7G, chamomile and marketed chamomile tea products at 366 nm. Track 1: A7G (25 ng), Track 2: wild chamomile, Tracks 3–7: M1-5, Tracks 8–18: A1-11; applied sample test solutions: 2 L/band; derivatization: NP/PEG 400.
other hand, hydrolysis caused to eliminate many other compounds which may help both to discriminate the genuine specimen and to evaluate the possible adulteration. It should be underlined that although A7G is an active marker, it should not be considered as a chemotaxonomic marker for chamomile authentication. Therefore, it is important to evaluate not only A7G but also other components for authentication of genuine chamomile. Consequently the method described for sample preparation in the present study may be considered as an ideal way for further authentication studies.
The presence of A7G was confirmed in all tea bag samples (M15) and bulk or packaged crude flowers sold as ‘chamomile’ except in A4, A9 and A10 (Table 3). M1-5 had almost identical densitograms as wild and reference chamomiles, while different fingerprint profiles were obtained for A1-11 samples, and none of them matched either with wild or reference chamomile fingerprints at 340 nm. These results demonstrated that crude flowers sold as chamomile in the spice shops were not genuine specimen, M. recutita. In addition, it was clearly shown that the densitograms of A1, A6 and A8;
40
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Fig. 2. HPTLC NH2 F254 s chromatogram of chamomile cultivars at 366 nm. Track 1: A7G (25 ng), Tracks 2a, b and c: chamomile (Bodegold), Tracks 3a, b and c: chamomile (Bona), Tracks 4a, b and c: chamomile (Zloyt Lan), Tracks 5a, b and c: chamomile (Yalova); arrangement of the samples on the plate according to their sowing dates starting by early November (a), early December (b) and late December (c), respectively; applied sample test solutions: 2 L/band; derivatization: NP/PEG 400.
Fig. 3. HPLC chromatogram of A7G standard solution (12.5 g/mL) and chamomile sample test solution (4 mg/mL) at 340 nm; 1: A7G in the standard solution; 1 : A7G in the sample test solution.
Fig. 4. HPTLC chromatogram captured at 366 nm (a) and densitograms at 340 nm (b, c) of chamomile extract. a1 and b: before ester hydrolysis; a2 and c: after ester hydrolysis; derivatization: NP/PEG 400.
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
41
Fig. 5. HPLC chromatograms of sample test solutions (4 mg/mL) at 340 nm.
A4 and A9; and A5 and A11 were found to be identical, respectively. Although each sample was purchased from different spice shops, these samples could be provided from the same suppliers. In our previous work, chromatograms of different species captured at 366 nm revealed some specific zones for different species: intense blue bright zones between RF = 0.42-0.5 for the flowers of Anthemis spp., Bellis spp., Tanacetum sp.; and additional intense orange zones close to RF = 0.1–0.16 and RF = 0.25–0.35 for the flowers of A. altissima L., A.tinctoria L. var. discoidea (All) D.C and A. scariosa L. [17]. These specific zones can be also seen in the chromatograms of A2-3, A5, and A7-10 (Fig. 1) and may eventually be an indication of adulteration of chamomile flowers with the species described above. A7G content determined by HPTLC method in different packed tea brands was ranged from 0.43 to 0.80 mg/g (Table 2), while in most of the crude chamomile samples sold in spice shops, A7G was either not determined or was below the limit of quantification (LOQ), which is 5 ng/band. Eventually, crude flowers sold as chamomile in spice shops would not met with therapeutic expectations of people. On the other hand, A7G content in A2 possibly adulterated with other Asteraceae flowers was found to be 0.51 mg/g, which is similar to M1-2 and M4 (Table 2). HPTLC densitograms of chamomile cultivar varieties (Bona, Bodegold and Zloyt Lan) and one cultivar from local population (Yalova) were also compared with wild and the reference chamomile samples and found to have identical fingerprint profile at 340 nm (Table 3). Besides, chemical profiles of chamomile
tea bags from different brands were found to be identical with those of cultivars at 340 nm (Table 3). However, A7G content in chamomile cultivars was higher than in other samples ranging from 2.33 to 4.14 mg/g (Table 2), which demonstrates that higher active ingredient content may be achieved through cultivation, controlled growing conditions. None of the crude flowers samples had HPTLC densitogram similar to chamomile cultivars at 340 nm. Densitograms comparison revealed that marketed chamomile tea products indicated that not chamomile tea bags but crude flowers sold on market were adulterated with other plant materials. The fact that chamomile is one of the most adulterated herbal drug with the other species of Asteraceae due to morphological similarities was described also by Joharchi and Amiri [24]. The application of recently reported HPTLC method for identification and quantification of some flavonoids such as A7G in methanolic extracts of cultivar chamomiles and marketed chamomile tea products [25] resulted in 32-fold difference between the lowest and highest amount of A7G (from 0.19 mg/g to 6.10 mg/g) in the analyzed tea bags, while the A7G content in chamomile cultivars varied from 1 to 2.83 mg/g. In another study of methanolic extracts of chamomile cultivars and its marketed tea products such as capsules, liquids and tea bags A7G content determined by HPLC-UV-MS was from 0.02 mg/g to 3.51 mg/g in chamomile cultivars and 0.08 mg/g to 2.45 mg/g in tea bags from different brands [26]. Consequently, it is obvious that the quality control of chamomile tea products is an inevitable task before marketing.
42
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Fig. 6. The line profile plots of HPTLC images for blue and green channels.
Moreover, M1-5 and randomly selected adulterated samples A1-2 and A8 were also analyzed by HPLC method described in the Ph. Eur. As a result, total A7G percentage in M1-5 was found as 0.15, 0.19, 0.27, 0.18 and 0.16, respectively. Results indicate that only one sample encoded M3 met the criterion (total A7G content should be
higher than 0.25%) of the Ph. Eur. When HPLC chromatograms of investigated samples and the reference chamomile (Fig. 3) were compared, it was observed that not only M1-5 samples but also A2 had chromatograms similar to that of the reference chamomile (Fig. 5). Since, A2 was found to be adulterated in the HPTLC analy-
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
43
Fig. 7. Principal component analysis of data from the blue channel (a) mutual projections of factor scores, (b) dendrogram, (c) and (d) loadings for the PC1 and PC2, respectively.
sis, this result indicates that HPTLC chemical fingerprinting profile was found to be superior to HPLC chemical fingerprinting profile to reveal adulteration. That could be due to ester hydrolysis step applied according to the methodology which resulted in conversion of many chemical markers in the sample (Fig. 4), which highlights that the sample preparation is one of the most important parts in plant analysis. Chamomile tea products are mostly marketed in the powdered form and therefore performing macroscopic analysis is not applicable. Moreover, due to similar microscopic characteristics of Asteraceae members, chamomile-like species cannot easily be discriminated from each other [27]. Although, a TLC method is stated in the Ph. Eur. monograph of chamomile, this is only for detection of the main components of its essential oil. Therefore, a simple, fast and selective method as described in this study should be considered before performing all long steps of HPLC before quantification of total A7G. 3.2. Fingerprint profiles according to image processing Pattern recognition techniques such as PCA and HCA were performed for the data set obtained by image analysis of HPTLC chromatograms using previously described procedure based on separated processing of blue, red and green channels, respectively [15,16]. Blue and green channels (Fig. 6) but not red channel showed well separation and provided good discrimination among samples and were therefore selected for further multivariate analysis. PCA was used to find the variables that are more informative to distinguish the samples, to reduce a multidimensional data set into
2 or 3 dimensions, and to recognize outliers. PCA creates a set of orthogonal axes variables called principal components (PCs) (linear combinations of the original variables) such that the first PC retains the maximum variation among the data [28]. Factor loading is the correlation between original variable and the PCs derived from PCA. Loading plot represents contribution of compounds from investigated samples to the total variability. The first two rotated factors (PCs) had the highest Eigen values and accounted for 39.20% and 18.10% of the total variability, respectively. Besides, first four components described 76.16% of total variability. B. perennis and one chamomile cultivar from Yalova flora were removed from the data set as outliers. According to PC1, samples were clustered into three groups, which mainly correspond to chamomile-like flowers from Asteraceae, different brands of chamomile tea bags and several sorts of chamomiles (Fig. 7a). Chamomile tea bags from different brands formed a cluster close to wild chamomile samples and Bellis sp., showing a similar chemical composition. Further, samples of crude flowers from different spice shops were overlapped with Anthemis spp., Tanacetum sp. and Chrysanthemum sp. Therefore, it is obvious that most of the crude chamomile flowers on the market might possibly be adulterated with these species. Additionally, score plot of PC1-PC2 showed a separated cluster consisted of three different varieties of chamomile cultivars (Bona, Bodegold and Zloyt Lan) and one cultivar from Yalova (Fig. 7a). Zones at RF = 0.37 (the zone with RF = 0.37 ± 0.01 identified as A7G according to the HPTLC chromatogram), RF = 0.40 and RF = 0.43 significantly contribute to differentiation along the PC1, while zone at RF = 0.79 showed the highest negative impact along the PC1
44
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Fig. 8. Principal component analysis of data from the green channel (a) mutual projections of factor scores, (b) dendrogram, (c) and (d) loadings for the PC1 and PC2, respectively.
direction and clearly differentiate chamomile cultivars from other samples. These samples had high peak at RF = 0.79 in blue channel profile. In addition, PC2 was highly positively contributed by zone at RF = 0.19 separating chamomile tea bags and wild chamomiles from other samples, and also highly negatively contributed by zones at RF = 0.43 and RF = 0.79, which differentiate samples of crude flowers and wild chamomiles (Fig. 7c and d). In order to search for natural groupings among the samples, cluster analysis was applied on data set. According to blue channel, dendrogram shows three main clusters. First cluster consisted of different brands of chamomile tea bags, wild chamomiles, Bodegold and one Bona type cultivars show similar profile. Second cluster consisted of crude flowers from spice shops grouped with Anthemis sp., Tanacetum sp. and Chrysanthemum sp., indicating their similar chemical profile resulted in either potential adulteration or misidentification. Bona and Zloyt Lan varieties of chamomile cultivars and one cultivar from Yalova formed a third cluster (Fig. 7b). In case of green channel, first principal component PC1 describes 38.48%, while PC2 19.69% of total variability. B. perennis was removed from data set as outlier. According to PCs score, crude flowers on market were overlapped with Anthemis sp., Tanacetum sp., and Chrysanthemum sp. indicating that probably crude flowers were adulterated with those species from Asteraceae. Chamomile tea bags from different brands were close to wild chamomile samples and Bellis sp., except several Anthemis spp. which formed separate cluster according to PC2. Chamomile cultivars were posi-
tioned on the upper right side indicating high dissimilarity from other Asteraceae samples (Fig. 8a). According to Fig. 8c, zones at RF = 0.28 and RF = 0.79 have positive correlation with PC1, while zones at RF = 0.37 and RF = 0.43 have the most negative impact on PC1, separating crude flowers and wild chamomiles. Zones at RF = 0.19, RF = 0.38, RF = 0.43 and RF = 0.79 significantly contribute to PC2 and differentiate crude flowers from tea bags and Anthemis sp. (Fig. 8d). In the HCA dendrogram, the 41 samples were clustered into three groups at a similarity level of 25% according to their HPTLC fingerprint profiles. The most of different brands of chamomile tea bags formed one cluster together with wild chamomile and two Anthemis spp. what is in agreement with previously obtained results. Crude chamomile flowers formed second cluster overlapped with Chrysanthemum sp, Anthemis sp., Bellis sp. and Tanacetum sp., while third cluster were composed of chamomile cultivars with different pattern (Fig. 8b). These results were also in agreement with those obtained previously. 4. Conclusion It is well-known that quality assurance of plant materials has a direct impact on the safety and efficacy. Herbal drugs should fulfill high demands concerning quality parameters before marketing. In this study, chamomile tea bags from different brands showed identical densitograms at 340 nm as wild and cultivar chamomiles, while such matching was not observed in case of densitograms of
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
crude flowers sold in spice shops. Therefore, it was concluded that crude flowers sold in spice shops were found to be adulterated. Besides, A7G content in different tea brands was ranged from 0.43 to 0.80 mg/g, while in most of the crude flowers A7G was neither determined nor below the LOQ. HPTLC fingerprinting profiles at 366 nm combined with multivariate image analysis and pattern recognition methods used for classification of wild and cultivar chamomiles as well as chamomile-like flowers from Asteraceae confirmed similarity between different chamomile tea bag brands and wild chamomiles, whereas crude flowers sold as chamomile in spice shops were found to be as similar as Anthemis spp. which indicated adulteration. Eventually, it is obvious that determination of not only the amount of active ingredient(s) but also ensuring genuine specimen plays a vital role for public health in order to avoid adverse reactions. Therefore, analysis of crude chamomile materials before marketing should be performed seriously for the quality assessment of commercial products. Acknowledgements Etil Guzelmeric expresses her sincere thanks to The Scientific and Technological Research Council of Turkey (TUBITAK) for the scholarship provided during Ph.D program. She also wishes to thank Dogan Arslan (Siirt University, Faculty of Agriculture, Department of Field Crops) and Ekrem Sezik (Faculty of Pharmacy, Yeditepe University, Istanbul, Turkey) for supplying cultivar chamomiles. References [1] M. Das, Chamomile: Medicinal Biomedical and Agricultural Aspects, CRC Press, Boca Raton, 2015. [2] D.L. McKay, J.B. Blumberg, Review article: a review of the bioactivity and potential health benefits of chamomile tea (Matricaria recutita L.), Phytother. Res. 20 (2006) 519–530. [3] J.K. Srivastava, S. Gupta, Extraction characterization, stability and biological activity of flavonoids isolated from chamomile flowers, Mol. Cell. Pharmacol. 1 (2009) 138–147. [4] J. Zhang, B. Wider, H. Shang, X. Li, E. Ernst, Quality of herbal medicines: challenges and Solutions, Complement. Ther. Med. 20 (2012) 100–106. [5] S¸. Kültür, Medicinal plants used in Kırklareli province (Turkey), J. Ethnopharmacol. 111 (2007) 341–364. [6] U. Cakilcioglu, S. Khatun, I. Turkoglu, S. Hayta, Ethnopharmacological survey of medicinal plants in Maden (Elazig-Turkey), J. Ethnopharmacol. 137 (2011) 469–486. [7] B. Özüdo˘gru, G. Akaydın, S. Erik, E. Yesilada, Inferences from an ethnobotanical field expedition in the selected locations of Sivas and Yozgat provinces (Turkey), J. Ethnopharmacol. 137 (2011) 85–98. [8] F. Tetik, S. Civelek, U. Cakilcioglu, Traditional uses of some medicinal plants in Malatya (Turkey), J. Ethnopharmacol. 146 (2013) 331–346. [9] B. Gürdal, S¸. Kültür, An ethnobotanical study of medicinal plants in Marmaris (Mu˘gla, Turkey), J. Ethnopharmacol. 146 (2013) 113–126.
45
[10] I. Kaval, L. Behc¸et, U. Cakilcioglu, Ethnobotanical study on medicinal plants in Gec¸itli and its surrounding (Hakkari-Turkey), J. Ethnopharmacol. 155 (2014) 171–184. [11] C. Bunchorntavakul, K.R. Reddy, Review article: herbal and dietary supplement hepatotoxicity, Aliment. Pharmacol. Ther. 37 (2013) 3–17. [12] S. Petronilho, M. Maraschin, M.A. Coimbra, S.M. Rocha, In vitro and in vivo studies of natural products: a challenge for their valuation. The case study of chamomile (Matricaria recutita L.), Ind. Crops Prod. 40 (2012) 1–12. [13] European pharmacopoeia (Ph. Eur.), Seventh, Council of Europe Strasbourg, France, (2010). ´ ´ F. Andric, ´ J. Trifkovic, ´ Planar P. Ristivojevic, [14] D. Milojkovic-Opsenica, chromatographic systems in pattern recognitionand fingerprint analysis, Chromatographia 76 (2013) 1239–1247. ´ F. Lj. Andric, ´ J.ð. Trifkovic, ´ I. Vovk, L.Zˇ . Stanisavljevic, ´ Zˇ .Lj. Teˇsic, ´ [15] P. Ristivojevic, ´ Pattern recognition methods and multivariate image D. Milojkovic-Opsenica, analysis in HPTLC fingerprinting of propolis extracts, J. Chemometr. 28 (2014) 301–306. ´ ´ J. Trifkovic, ´ I. Vovk, D. Luˇsic, ´ Zˇ . Teˇsic, ´ [16] D. Milojkovic-Opsenica, P. Ristivojevic, TLC fingerprinting and pattern recognition methods in the assessment of authenticity of poplar-type propolis, J. Chromatogr. Sci. (2016) 1–7. [17] E. Guzelmeric, I. Vovk, E. Yesilada, Development and validation of an HPTLC method for apigenin 7-O-glucoside in chamomile flowers and its application for fingerprint discrimination of chamomile-like materials, J. Pharm. Biomed. Anal. 107 (2015) 108–118. [18] E. Reich, A. Schibli, High Performance Thin Layer Chromatography for the Analysis of Medicinal Plants, first ed., Thieme Stuttgart, New York, 2007. [19] Eigenvector Research, Inc Retrived March 15, 2015 from: http://www. eigenvector.com/software/pls toolbox.htm. [20] K.H. Wong, V. Razmovski-Naumovski, K.M. Li, G.Q. Li, K. Chan, Differentiating puerariae lobatae radix and puerariae thomsonii radix using HPTLC coupled with multivariate classification analyses, J. Pharm. Biomed. Anal. 95 (2014) 11–19. [21] N. Harbourne, J.C. Jacquier, D. O’Riordan, Optimisation of the extraction and processing conditions of chamomile (Matricaria chamomilla L.) for incorporation into a beverage, Food Chem. 115 (2009) 15–19. [22] F.N. Fonseca, M.F.M. Tavares, Validation of a capillary electrophoresis method for the quantitative determination of free and total apigenin in extracts of Chamomilla recutita, Phytochem. Anal. 15 (2004) 65–70. [23] G. Haghi, A. Hatami, A. Safaei, M. Mehran, Analysis of phenolic compounds in Matricaria chamomilla and its extracts by UPLC-UV, Res. Pharm. Sci. 9 (2014) 31–37. [24] M.R. Joharchi, M.S. Amiri, Taxanomic evaluation of misidentification of crude herbal drugs marketed in Iran, Avicenna J. Phytomed. 2 (2012) 105–112. [25] S. Sagi, B. Avula, Y.H. Wang, J. Zhao, I.A. Khan, Quantitative determination of seven chemical constituents and chemo-type differentiation of chamomiles using high-performance thin-layer chromatography, J. Sep. Sci. 37 (2014) 2797–2804. [26] B. Avula, Y.H. Wang, M. Wang, C. Avonto, J. Zhao, T.J. Smillie, D. Rua, I.A. Khan, Quantitative determination of phenolic compounds by UHPLC-UV-MS and use of partial least-square discriminant analysis to differentiate chemo-types of Chamomile/Chrysanthemum flower heads, J. Pharm. Biomed. Anal. 88 (2014) 278–288. [27] D. C¸alıs¸kan, Papatya Olarak Satılan Bitkiler Üzerinde Fitoterapötik C¸alıs¸malar. Master Thesis, Gazi University, Ankara, 2010. [28] L. Giansante, D. Di Vincenzo, G. Bianchi, Classification of monovarietal Italian olive oils by unsupervised (PCA) and supervised (LDA) chemometrics, J. Sci. Food Agric. 83 (2003) 905–911.
Contents lists available at ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba
Quality assessment of marketed chamomile tea products by a validated HPTLC method combined with multivariate analysis d ´ Etil Guzelmeric a , Petar Ristivojevic´ b , Irena Vovk c , Duˇsanka Milojkovic-Opsenica , a,∗ Erdem Yesilada a
Yeditepe University, Faculty of Pharmacy, Department of Pharmacognosy and Phytotherapy, Kayisdagi Cad., Atasehir, 34755 Istanbul, Turkey Innovation Centre of Faculty of Chemistry Ltd., Studenski trg 12-16, 11000 Belgrade, Serbia c Department of Food Chemistry, National Institute of Chemistry, Hajdrihova 19, SI-1000 Ljubljana, Slovenia d Faculty of Chemistry, University of Belgrade, Studentski trg 12-16, 11158 Belgrade, Serbia b
a r t i c l e
i n f o
Article history: Received 15 June 2016 Received in revised form 19 September 2016 Accepted 24 September 2016 Available online 26 September 2016 Keywords: Matricaria recutita L. High performance thin-layer chromatography (HPTLC) Apigenin 7-O-glucoside Quality control Chamomile tea products Multivariate analysis
a b s t r a c t Chamomile tea composed of dried flower heads of Matricaria recutita L. (Asteraceae) is one of the most popular single ingredient herbal teas. Tea industries, spice shops or public bazaars are mostly supplied chamomile as a raw material via cultivation or through nature-picking. However, one of the drawbacks of nature-picking is adulteration. This could be either due to false authentication of the plant materials by ingenuous pickers or intentional/unintentional substitution with other flowers resembling to chamomile in appearance during harvesting. Therefore, quality control of raw chamomile materials before marketing should be carefully considered not only by quantification of apigenin 7-O-glucoside (active marker) but also by fingerprinting of chemical composition. This work presents both quantification of apigenin 7-Oglucoside and chemical fingerprinting of commercial chamomile tea products obtained from different food stores and spice shops by a validated HPTLC method. In addition, HPTLC profiles of investigated chamomile tea samples were compared with HPLC method stated in the European Pharmacopoeia and it was found that HPTLC method was superior to HPLC method in the field of adulteration confirmation. Therefore, fingerprint profiles performed on the silica gel 60 NH2 F254 s HPTLC plates combined with pattern recognition techniques of these marketed products were comparatively evaluated with wild and cultivar chamomile samples and also chamomile-like species from Asteraceae. Consequently, not chamomile tea bags but crude flowers sold on market were found to be adulterated with other plant materials. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Brewed tea of chamomile flowers (Matricaria recutita L., Asteraceae) has been consumed for centuries either for its delight or healing benefits such as alleviation of the symptoms of gastrointestinal complaints, inflammatory and nervous problems. Apart from its traditional use, recent investigations on health promoting efficacy of chamomile tea have also revealed its antioxidant, antinociceptive, anticancer etc. activities mostly associated with one of its major components, apigenin 7-O-glucoside (A7G) [1,2]. Today, daily more than one million cups of chamomile tea have been consumed throughout the world [3]. Tea industries, food stores, spice shops or public bazaars are mostly supplied
∗ Corresponding author. E-mail address: [email protected] (E. Yesilada). http://dx.doi.org/10.1016/j.jpba.2016.09.030 0731-7085/© 2016 Elsevier B.V. All rights reserved.
chamomile flowers as a raw material in two ways: cultivation or through nature-picking. Then, they introduce them into market both crude (bulk or wrapped) and processed (tea bags). Actually, retailers commonly prefer nature-picking materials due to its lower cost. However, one of the drawbacks of such materials is adulteration due to false authentication of the plant materials by ingenuous pickers or intentional/unintentional substitution with other flowers resembling to chamomile in appearance during harvesting [4]. On the other hand, recent ethnobotanical surveys carried out in different localities of Turkey revealed that several other species from Asteraceae with similar appearance are also named “papatya” [Turkish name of chamomile] i.e. Anthemis tinctoria L., Anthemis cretica L., Anthemis cotula L., Anthemis austriaca L., Anthemis ceolopoda L., Anthemis wiedemanniana L., Bellis perennis L., Chrysanthemum coronarium L. [5–10]. Even more important point, severe problems may be emerged by the use of such false chamomile materials i.e. Senecio sp., which is reported to induce
36
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Table 1 Analyzed chamomile samples in this study. Number of the samples
Samples
Sample Codes
Detailed Information
5 11 14
Tea bags Bulk or packaged crude flowers sold as ‘chamomile’ Wild chamomiles and chamomile-like flowers belonging to 4 different genera, Anthemis L., Bellis L., Tanacetum L. and Chrysanthemum L. Bona, Bodegold and Zloyt Lan varieties of chamomile cultivars and one cultivar from local population (Yalova) having 3 different sowing dates (early November, early December and late December)
M1-5 A1-11 –
purchased from food stores (Istanbul, Turkey) purchased from bazaar (Istanbul, Turkey) gathered from nature in different localities of Turkey by local people assuming as chamomile.
–
obtained from Ataturk Central Horticultural Research Institute (Yalova, Turkey)
12
severe hepatic veno-occlusive disease and even death [11]. Moreover, loss in the expected efficacy might be the other negative impact due to low yield or absence of active ingredient in the false materials. The widespread usage and well-known popularity of chamomile flowers have raisen it to be documented in the pharmacopoeias of 26 countries including European Pharmacopoeia (Ph. Eur.) [12]. In the Ph. Eur., A7G is stated as an active marker of chamomile flowers and an high-performance liquid chromatographic (HPLC) method is described for estimation of its total concentration [13]. High performance thin-layer chromatography (HPTLC) as an analytical technique provides convenience for both qualitative and quantitative conclusions. Nowadays, HPTLC fingerprints in combination with pattern recognition methods has become popular either to provide useful information on complex matrices or to support analytical results [14–16]. The aim of the present study was to: 1.) reveal chemical fingerprint profiles of the marketed chamomile tea products and quantify A7G in these samples using previously reported validated HPTLC method [17]; 2.) compare the selectivity of the HPTLC method and HPLC method stated in the Ph. Eur. in discrimination of the genuine specimen; 3.) analyze and evaluate HPTLC fingerprint profiles of marketed chamomile tea products with those of wild and cultivar chamomile samples, as well as other species having chamomile-like flowers (Anthemis spp., Bellis spp., Tanacetum sp., Chrysanthemum sp.) from the family Asteraceae using pattern recognition techniques.
2. Experimental
Table 2 Commercial chamomile tea products and chamomile cultivars: yields (%) after lyophilization and A7G contents.
Tea products
Cultivars
Product
Yield (%)
A7G (mg/g)x
M1 M2 M3 M4 M5 A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 Bodegolda Bodegoldb Bodegoldc Bonaa Bonab Bonac Zloyt Lana Zloyt Lanb Zloyt Lanc Yalovaa Yalovab Yalovac
17.49 20.24 24.10 19.82 28.37 24.07 28.63 19.14 20.35 14.98 15.97 17.64 21.62 20.04 12.70 13.84 34.05 32.47 29.32 30.07 31.09 29.40 29.83 32.25 31.8 42.05 39.19 36.05
0.49 ± 0.01 0.49 ± 0.03 0.80 ± 0.02 0.54 ± 0.01 0.43 ± 0.01 0.30 ± 0.02 0.51 ± 0.01
ndt: A7G not detected; < LOQ: under the limit of quantitation (5 ng/band). a,b,c Sowing dates starting by early November (a), early December (b) and late December (c), respectively. x Mean ± SD (n = 3).
2.1. Chemicals HPLC grade acetonitrile was obtained from J. T. Baker (Deventer, the Netherlands). The other solvents were of analytical grade. Ethanol absolute, ethyl acetate and formic acid were purchased from Sigma-Aldrich (Steinheim, Germany); acetic acid was from Riedel-de Haen (Seelze, Germany), o-phosphoric acid was from Merck (Darmstadt, Germany); methanol and dichloromethane were from Analar Normapur (Muarrie, Australia). 2-aminoethyl diphenylborinate and polyethylene glycol 400 were from Fluka (Steinheim, Germany) and Merck (Hohenbrunn, Germany), respectively. The ultrapure water was obtained from Millipore, Simplicity UV (Darmstadt, Germany). Sodium hydroxide, magnesium chloride-6-hydrate as well as standards of A7G and 5,7-dihydroxy-4-methylcoumarin were purchased from Sigma-Aldrich (Steinheim, Germany).
2.2. Plant materials The analyzed samples were chamomile tea bags, crude chamomile flowers, wild and cultivar chamomiles as well as several chamomile-like flowers from Asteraceae (Table 1).
The materials stored in a refrigerator at −25◦ C were ground to powder in a mechanic grinder before extraction. In addition, six tea bags from each different tea brands (M1-5) being in a powdered form were randomly selected due to variable weight (1.3–2.0 g) and mixed in a beaker for homogenous sampling.
2.3. Preparation of standard solutions 2.3.1. Standard solutions for HPTLC analysis A7G stock solution (250 g/mL) was prepared in methanol and further diluted with the same solvent to prepare the working solutions 2.5, 5 and 10 g/mL.
2.3.2. Standard solutions for HPLC analysis A7G standard solution (12.5 g/mL) and standard solution mixture containing both A7G and 5,7-dihydroxy-4-methylcoumarin (10 g/mL) were prepared according to the procedure described in the Ph. Eur. [13].
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
2.4. Preparation of sample test solutions 2.4.1. Extraction of wild and cultivar chamomiles, chamomile-like flowers and chamomile tea products for HPTLC analysis All samples were brewed to prepare 2% infusion following our previously described procedure [17]. Next, they were filtered, cooled and finally freeze-dried. The yields of the lyophilized commercial and cultivar samples are listed in Table 2. The yields of other samples are stated in the previous report [17]. Then, 10 mg of each wild chamomile and chamomile-like sample and 20 mg of each chamomile tea product was accurately weighed and extracted with 10 mL of methanol in an ultrasonic bath for 15 min. Suspended particles were removed by filtration through a 0.45 m RC-membrane filter (Sartorius stedim biotech). Further, the sample test solutions of chamomile cultivars were diluted 2 times.
37
HPLC system suitability confirmed by the resolution (Rs ) between the standards A7G and 5,7-dihydroxy-4-methylcoumarin which was found as 7.54 ± 0.03 (n = 3). The result was found to be suitable for further analyses according to the Ph. Eur. The retention time (tR ) of the A7G standard (tR = 6.35 ± 0.05; n = 3), presented in Fig. 3 were compared with the tR obtained for the reference chamomile sample solution to evaluate the identity of the A7G. 2.7. Statistical analysis The quantitative results of HPTLC analyses of each sample test solution applied in triplicate on the same plate were expressed as mean ± standard deviation (SD). These treatments were carried out by using SPSS Data Editor (version 20.0). 2.8. Image processing and multivariate analysis
2.4.2. Extraction of reference chamomile and chamomile tea products for HPLC analysis One gram of the powdered chamomile flowers were accurately weighed and extracted with 100 mL EtOH (96%), and then heated under a reflux condenser on a water-bath for 15 min and cooled and filtered with a filter paper. 10 mL of freshly prepared diluted sodium hydroxide solution was added to the filtrate and the mixture was heated under a reflux condenser on a water-bath for 1 h and cooled. It was diluted to 125 mL with EtOH (96%). To 25 mL of the solution, 0.25 g of citric acid was added and shook for 5 min. First, it was filtered with a filter paper and then filtered through a 0.45 m RCmembrane filter. 5 mL of this solution diluted to 10 mL with the initial mixture of the mobile phase [13]. 2.5. HPTLC method All samples were analyzed by previously validated HPTLC method [17]. HPTLC glass plates (20 cm × 10 cm) precoated with silica gel 60 NH2 F254 s (Merck), Linomat V automatic sample spotter (Camag, Muttenz, Switzerland) equipped with 100 L Hamilton syringe (Bonaduz, Switzerland), Automatic Developing Chamber 2 (ADC2, Camag) and TLC Scanner 3 (Camag) were used during analyses. After development with ethyl acetate-formic acid-acetic acid-water (30:1.5:1.5:3, v/v/v/v), the plates were heated at 100◦ C on the Camag TLC plate heater for 3 min and dipped into Natural Products (NP) and Polyethylene glycol (PEG) 400 solutions prepared as stated by Reich and Schibli [18]. After, the plates were documented by the Camag TLC visualizer at 366 nm. All the instruments were operated by winCATS program (version 1.4.8, Camag). The identity of the A7G in all samples was evaluated by comparison of the retention factors (RF ) of the sample zone and the zone of A7G standard, 0.38 ± 0.01, (Figs. 1 and 2). For the quantitative analysis, 5, 10, 15, 20, 25 and 50 ng/band A7G standard solution was applied on the plate for the calibration curve and A7G content was evaluated in all samples through peak area via quadratic regression (R2 = 0.9995). The application volume of each sample test solutions was 2 L. 2.6. HPLC method HPLC analyses were performed as described in the Ph. Eur. monograph for chamomile [13] and carried out by using the following equipments; Agilent 1100 HPLC system (Technologies, Santa Clara, California, USA) coupled with ChemStation 10.01 software, a model G1379A vacuum degasser, a model G1311A quaternary pump, a model G1313A auto-sampler, a model G1316A thermostated column compartment, and a model G1315 B diode array detector. Separations were achieved on an Agilent Zorbax Eclipse Plus C18 ODS column (250 mm × 4.6 mm, 5 m particle size).
Images of the HPTLC chromatograms were transformed to data set which contains 42 samples per 459 pixels for red, green and blue channels. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) as pattern recognition techniques were performed by using PLS Toolbox, v.6.2.1, for MATLAB 7.12.0 (R2011a) [19]. PCA was performed by using a singular value decomposition algorithm and a 0.95 confidence level for Q and T2 Hostelling limits for outliers. HCA was obtained by using the ward method to calculate cluster distances and applying Euclidean distance as a measure of distance between the samples. All data obtained from HPTLC chromatograms were preprocessed by using mean centered, standard normal variety (SNV) prior to any multivariate analysis. Correlation optimised warping (COW) was applied to correct the inter- and intra-plate peak shift due to variations in experimental conditions, analyst error and instrumental instability [20]. 3. Results and discussion 3.1. Quality assessment of marketed chamomile tea products For centuries, chamomile tea has been consumed due to either its pleasant taste or medicinal purposes. In the light of the foregoing, all samples were prepared by brewing at 100◦ C. Recently, Harbourne et al. investigated the A7G content of chamomile teas at different preparation temparatures through steeping in 100 mL of distilled water at 57, 70, 80, 90 and 100◦ C. They reported that the A7G concentration incrementally increased between 57 and 90◦ C and its content became stable from 90◦ C to 100◦ C [21]. Their optimised extraction procedure yielding the highest A7G content was found to be in accordance with the results of the present study. Moreover, the preferred extraction in this study was also compared with described sample preparation for HPLC analysis stated in the Ph. Eur. [13]. It should be noted that in the Ph. Eur. instead of free A7G content, total A7G amount (including monoor diacetylated derivatives of A7G) is estimated in hydroalcoholic chamomile extract. Therefore, after extraction of powdered chamomile flowers with 96% ethanol, the extract was subjected to HPTLC analysis both directly and following the ester hydrolysis by addition sodium hydroxide reagent to the extract and heating under a reflux condenser on a water-bath for 1 h. According to the HPTLC chromatogram in Fig. 4a (Track 1), many components disappeared after addition of sodium hydroxide possibly due to ester hydrolysis, Fig. 4a (Track 2). Moreover, the zones around RF = 0.60 and RF = 0.70 having characteristics color similar to that of A7G in Fig. 4a (Track 1) were not seen after hydrolysis in Figs. 4b and 4c; these zones could be mono- or diacetylated A7G derivatives. Fonseca and Tavares and Haghi et al. also pointed out that following ester hydrolysis, the peak area of A7G expanded [22,23]. On the
38
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Table 3 HPTLC densitograms of commercial chamomile tea products and chamomile cultivars.
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
39
Table 3 (Continued)
a,b,c
Sowing dates starting by early November (a), early December (b) and late December (c), respectively.
Fig. 1. HPTLC NH2 F254 s chromatogram of A7G, chamomile and marketed chamomile tea products at 366 nm. Track 1: A7G (25 ng), Track 2: wild chamomile, Tracks 3–7: M1-5, Tracks 8–18: A1-11; applied sample test solutions: 2 L/band; derivatization: NP/PEG 400.
other hand, hydrolysis caused to eliminate many other compounds which may help both to discriminate the genuine specimen and to evaluate the possible adulteration. It should be underlined that although A7G is an active marker, it should not be considered as a chemotaxonomic marker for chamomile authentication. Therefore, it is important to evaluate not only A7G but also other components for authentication of genuine chamomile. Consequently the method described for sample preparation in the present study may be considered as an ideal way for further authentication studies.
The presence of A7G was confirmed in all tea bag samples (M15) and bulk or packaged crude flowers sold as ‘chamomile’ except in A4, A9 and A10 (Table 3). M1-5 had almost identical densitograms as wild and reference chamomiles, while different fingerprint profiles were obtained for A1-11 samples, and none of them matched either with wild or reference chamomile fingerprints at 340 nm. These results demonstrated that crude flowers sold as chamomile in the spice shops were not genuine specimen, M. recutita. In addition, it was clearly shown that the densitograms of A1, A6 and A8;
40
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Fig. 2. HPTLC NH2 F254 s chromatogram of chamomile cultivars at 366 nm. Track 1: A7G (25 ng), Tracks 2a, b and c: chamomile (Bodegold), Tracks 3a, b and c: chamomile (Bona), Tracks 4a, b and c: chamomile (Zloyt Lan), Tracks 5a, b and c: chamomile (Yalova); arrangement of the samples on the plate according to their sowing dates starting by early November (a), early December (b) and late December (c), respectively; applied sample test solutions: 2 L/band; derivatization: NP/PEG 400.
Fig. 3. HPLC chromatogram of A7G standard solution (12.5 g/mL) and chamomile sample test solution (4 mg/mL) at 340 nm; 1: A7G in the standard solution; 1 : A7G in the sample test solution.
Fig. 4. HPTLC chromatogram captured at 366 nm (a) and densitograms at 340 nm (b, c) of chamomile extract. a1 and b: before ester hydrolysis; a2 and c: after ester hydrolysis; derivatization: NP/PEG 400.
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
41
Fig. 5. HPLC chromatograms of sample test solutions (4 mg/mL) at 340 nm.
A4 and A9; and A5 and A11 were found to be identical, respectively. Although each sample was purchased from different spice shops, these samples could be provided from the same suppliers. In our previous work, chromatograms of different species captured at 366 nm revealed some specific zones for different species: intense blue bright zones between RF = 0.42-0.5 for the flowers of Anthemis spp., Bellis spp., Tanacetum sp.; and additional intense orange zones close to RF = 0.1–0.16 and RF = 0.25–0.35 for the flowers of A. altissima L., A.tinctoria L. var. discoidea (All) D.C and A. scariosa L. [17]. These specific zones can be also seen in the chromatograms of A2-3, A5, and A7-10 (Fig. 1) and may eventually be an indication of adulteration of chamomile flowers with the species described above. A7G content determined by HPTLC method in different packed tea brands was ranged from 0.43 to 0.80 mg/g (Table 2), while in most of the crude chamomile samples sold in spice shops, A7G was either not determined or was below the limit of quantification (LOQ), which is 5 ng/band. Eventually, crude flowers sold as chamomile in spice shops would not met with therapeutic expectations of people. On the other hand, A7G content in A2 possibly adulterated with other Asteraceae flowers was found to be 0.51 mg/g, which is similar to M1-2 and M4 (Table 2). HPTLC densitograms of chamomile cultivar varieties (Bona, Bodegold and Zloyt Lan) and one cultivar from local population (Yalova) were also compared with wild and the reference chamomile samples and found to have identical fingerprint profile at 340 nm (Table 3). Besides, chemical profiles of chamomile
tea bags from different brands were found to be identical with those of cultivars at 340 nm (Table 3). However, A7G content in chamomile cultivars was higher than in other samples ranging from 2.33 to 4.14 mg/g (Table 2), which demonstrates that higher active ingredient content may be achieved through cultivation, controlled growing conditions. None of the crude flowers samples had HPTLC densitogram similar to chamomile cultivars at 340 nm. Densitograms comparison revealed that marketed chamomile tea products indicated that not chamomile tea bags but crude flowers sold on market were adulterated with other plant materials. The fact that chamomile is one of the most adulterated herbal drug with the other species of Asteraceae due to morphological similarities was described also by Joharchi and Amiri [24]. The application of recently reported HPTLC method for identification and quantification of some flavonoids such as A7G in methanolic extracts of cultivar chamomiles and marketed chamomile tea products [25] resulted in 32-fold difference between the lowest and highest amount of A7G (from 0.19 mg/g to 6.10 mg/g) in the analyzed tea bags, while the A7G content in chamomile cultivars varied from 1 to 2.83 mg/g. In another study of methanolic extracts of chamomile cultivars and its marketed tea products such as capsules, liquids and tea bags A7G content determined by HPLC-UV-MS was from 0.02 mg/g to 3.51 mg/g in chamomile cultivars and 0.08 mg/g to 2.45 mg/g in tea bags from different brands [26]. Consequently, it is obvious that the quality control of chamomile tea products is an inevitable task before marketing.
42
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Fig. 6. The line profile plots of HPTLC images for blue and green channels.
Moreover, M1-5 and randomly selected adulterated samples A1-2 and A8 were also analyzed by HPLC method described in the Ph. Eur. As a result, total A7G percentage in M1-5 was found as 0.15, 0.19, 0.27, 0.18 and 0.16, respectively. Results indicate that only one sample encoded M3 met the criterion (total A7G content should be
higher than 0.25%) of the Ph. Eur. When HPLC chromatograms of investigated samples and the reference chamomile (Fig. 3) were compared, it was observed that not only M1-5 samples but also A2 had chromatograms similar to that of the reference chamomile (Fig. 5). Since, A2 was found to be adulterated in the HPTLC analy-
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
43
Fig. 7. Principal component analysis of data from the blue channel (a) mutual projections of factor scores, (b) dendrogram, (c) and (d) loadings for the PC1 and PC2, respectively.
sis, this result indicates that HPTLC chemical fingerprinting profile was found to be superior to HPLC chemical fingerprinting profile to reveal adulteration. That could be due to ester hydrolysis step applied according to the methodology which resulted in conversion of many chemical markers in the sample (Fig. 4), which highlights that the sample preparation is one of the most important parts in plant analysis. Chamomile tea products are mostly marketed in the powdered form and therefore performing macroscopic analysis is not applicable. Moreover, due to similar microscopic characteristics of Asteraceae members, chamomile-like species cannot easily be discriminated from each other [27]. Although, a TLC method is stated in the Ph. Eur. monograph of chamomile, this is only for detection of the main components of its essential oil. Therefore, a simple, fast and selective method as described in this study should be considered before performing all long steps of HPLC before quantification of total A7G. 3.2. Fingerprint profiles according to image processing Pattern recognition techniques such as PCA and HCA were performed for the data set obtained by image analysis of HPTLC chromatograms using previously described procedure based on separated processing of blue, red and green channels, respectively [15,16]. Blue and green channels (Fig. 6) but not red channel showed well separation and provided good discrimination among samples and were therefore selected for further multivariate analysis. PCA was used to find the variables that are more informative to distinguish the samples, to reduce a multidimensional data set into
2 or 3 dimensions, and to recognize outliers. PCA creates a set of orthogonal axes variables called principal components (PCs) (linear combinations of the original variables) such that the first PC retains the maximum variation among the data [28]. Factor loading is the correlation between original variable and the PCs derived from PCA. Loading plot represents contribution of compounds from investigated samples to the total variability. The first two rotated factors (PCs) had the highest Eigen values and accounted for 39.20% and 18.10% of the total variability, respectively. Besides, first four components described 76.16% of total variability. B. perennis and one chamomile cultivar from Yalova flora were removed from the data set as outliers. According to PC1, samples were clustered into three groups, which mainly correspond to chamomile-like flowers from Asteraceae, different brands of chamomile tea bags and several sorts of chamomiles (Fig. 7a). Chamomile tea bags from different brands formed a cluster close to wild chamomile samples and Bellis sp., showing a similar chemical composition. Further, samples of crude flowers from different spice shops were overlapped with Anthemis spp., Tanacetum sp. and Chrysanthemum sp. Therefore, it is obvious that most of the crude chamomile flowers on the market might possibly be adulterated with these species. Additionally, score plot of PC1-PC2 showed a separated cluster consisted of three different varieties of chamomile cultivars (Bona, Bodegold and Zloyt Lan) and one cultivar from Yalova (Fig. 7a). Zones at RF = 0.37 (the zone with RF = 0.37 ± 0.01 identified as A7G according to the HPTLC chromatogram), RF = 0.40 and RF = 0.43 significantly contribute to differentiation along the PC1, while zone at RF = 0.79 showed the highest negative impact along the PC1
44
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
Fig. 8. Principal component analysis of data from the green channel (a) mutual projections of factor scores, (b) dendrogram, (c) and (d) loadings for the PC1 and PC2, respectively.
direction and clearly differentiate chamomile cultivars from other samples. These samples had high peak at RF = 0.79 in blue channel profile. In addition, PC2 was highly positively contributed by zone at RF = 0.19 separating chamomile tea bags and wild chamomiles from other samples, and also highly negatively contributed by zones at RF = 0.43 and RF = 0.79, which differentiate samples of crude flowers and wild chamomiles (Fig. 7c and d). In order to search for natural groupings among the samples, cluster analysis was applied on data set. According to blue channel, dendrogram shows three main clusters. First cluster consisted of different brands of chamomile tea bags, wild chamomiles, Bodegold and one Bona type cultivars show similar profile. Second cluster consisted of crude flowers from spice shops grouped with Anthemis sp., Tanacetum sp. and Chrysanthemum sp., indicating their similar chemical profile resulted in either potential adulteration or misidentification. Bona and Zloyt Lan varieties of chamomile cultivars and one cultivar from Yalova formed a third cluster (Fig. 7b). In case of green channel, first principal component PC1 describes 38.48%, while PC2 19.69% of total variability. B. perennis was removed from data set as outlier. According to PCs score, crude flowers on market were overlapped with Anthemis sp., Tanacetum sp., and Chrysanthemum sp. indicating that probably crude flowers were adulterated with those species from Asteraceae. Chamomile tea bags from different brands were close to wild chamomile samples and Bellis sp., except several Anthemis spp. which formed separate cluster according to PC2. Chamomile cultivars were posi-
tioned on the upper right side indicating high dissimilarity from other Asteraceae samples (Fig. 8a). According to Fig. 8c, zones at RF = 0.28 and RF = 0.79 have positive correlation with PC1, while zones at RF = 0.37 and RF = 0.43 have the most negative impact on PC1, separating crude flowers and wild chamomiles. Zones at RF = 0.19, RF = 0.38, RF = 0.43 and RF = 0.79 significantly contribute to PC2 and differentiate crude flowers from tea bags and Anthemis sp. (Fig. 8d). In the HCA dendrogram, the 41 samples were clustered into three groups at a similarity level of 25% according to their HPTLC fingerprint profiles. The most of different brands of chamomile tea bags formed one cluster together with wild chamomile and two Anthemis spp. what is in agreement with previously obtained results. Crude chamomile flowers formed second cluster overlapped with Chrysanthemum sp, Anthemis sp., Bellis sp. and Tanacetum sp., while third cluster were composed of chamomile cultivars with different pattern (Fig. 8b). These results were also in agreement with those obtained previously. 4. Conclusion It is well-known that quality assurance of plant materials has a direct impact on the safety and efficacy. Herbal drugs should fulfill high demands concerning quality parameters before marketing. In this study, chamomile tea bags from different brands showed identical densitograms at 340 nm as wild and cultivar chamomiles, while such matching was not observed in case of densitograms of
E. Guzelmeric et al. / Journal of Pharmaceutical and Biomedical Analysis 132 (2017) 35–45
crude flowers sold in spice shops. Therefore, it was concluded that crude flowers sold in spice shops were found to be adulterated. Besides, A7G content in different tea brands was ranged from 0.43 to 0.80 mg/g, while in most of the crude flowers A7G was neither determined nor below the LOQ. HPTLC fingerprinting profiles at 366 nm combined with multivariate image analysis and pattern recognition methods used for classification of wild and cultivar chamomiles as well as chamomile-like flowers from Asteraceae confirmed similarity between different chamomile tea bag brands and wild chamomiles, whereas crude flowers sold as chamomile in spice shops were found to be as similar as Anthemis spp. which indicated adulteration. Eventually, it is obvious that determination of not only the amount of active ingredient(s) but also ensuring genuine specimen plays a vital role for public health in order to avoid adverse reactions. Therefore, analysis of crude chamomile materials before marketing should be performed seriously for the quality assessment of commercial products. Acknowledgements Etil Guzelmeric expresses her sincere thanks to The Scientific and Technological Research Council of Turkey (TUBITAK) for the scholarship provided during Ph.D program. She also wishes to thank Dogan Arslan (Siirt University, Faculty of Agriculture, Department of Field Crops) and Ekrem Sezik (Faculty of Pharmacy, Yeditepe University, Istanbul, Turkey) for supplying cultivar chamomiles. References [1] M. Das, Chamomile: Medicinal Biomedical and Agricultural Aspects, CRC Press, Boca Raton, 2015. [2] D.L. McKay, J.B. Blumberg, Review article: a review of the bioactivity and potential health benefits of chamomile tea (Matricaria recutita L.), Phytother. Res. 20 (2006) 519–530. [3] J.K. Srivastava, S. Gupta, Extraction characterization, stability and biological activity of flavonoids isolated from chamomile flowers, Mol. Cell. Pharmacol. 1 (2009) 138–147. [4] J. Zhang, B. Wider, H. Shang, X. Li, E. Ernst, Quality of herbal medicines: challenges and Solutions, Complement. Ther. Med. 20 (2012) 100–106. [5] S¸. Kültür, Medicinal plants used in Kırklareli province (Turkey), J. Ethnopharmacol. 111 (2007) 341–364. [6] U. Cakilcioglu, S. Khatun, I. Turkoglu, S. Hayta, Ethnopharmacological survey of medicinal plants in Maden (Elazig-Turkey), J. Ethnopharmacol. 137 (2011) 469–486. [7] B. Özüdo˘gru, G. Akaydın, S. Erik, E. Yesilada, Inferences from an ethnobotanical field expedition in the selected locations of Sivas and Yozgat provinces (Turkey), J. Ethnopharmacol. 137 (2011) 85–98. [8] F. Tetik, S. Civelek, U. Cakilcioglu, Traditional uses of some medicinal plants in Malatya (Turkey), J. Ethnopharmacol. 146 (2013) 331–346. [9] B. Gürdal, S¸. Kültür, An ethnobotanical study of medicinal plants in Marmaris (Mu˘gla, Turkey), J. Ethnopharmacol. 146 (2013) 113–126.
45
[10] I. Kaval, L. Behc¸et, U. Cakilcioglu, Ethnobotanical study on medicinal plants in Gec¸itli and its surrounding (Hakkari-Turkey), J. Ethnopharmacol. 155 (2014) 171–184. [11] C. Bunchorntavakul, K.R. Reddy, Review article: herbal and dietary supplement hepatotoxicity, Aliment. Pharmacol. Ther. 37 (2013) 3–17. [12] S. Petronilho, M. Maraschin, M.A. Coimbra, S.M. Rocha, In vitro and in vivo studies of natural products: a challenge for their valuation. The case study of chamomile (Matricaria recutita L.), Ind. Crops Prod. 40 (2012) 1–12. [13] European pharmacopoeia (Ph. Eur.), Seventh, Council of Europe Strasbourg, France, (2010). ´ ´ F. Andric, ´ J. Trifkovic, ´ Planar P. Ristivojevic, [14] D. Milojkovic-Opsenica, chromatographic systems in pattern recognitionand fingerprint analysis, Chromatographia 76 (2013) 1239–1247. ´ F. Lj. Andric, ´ J.ð. Trifkovic, ´ I. Vovk, L.Zˇ . Stanisavljevic, ´ Zˇ .Lj. Teˇsic, ´ [15] P. Ristivojevic, ´ Pattern recognition methods and multivariate image D. Milojkovic-Opsenica, analysis in HPTLC fingerprinting of propolis extracts, J. Chemometr. 28 (2014) 301–306. ´ ´ J. Trifkovic, ´ I. Vovk, D. Luˇsic, ´ Zˇ . Teˇsic, ´ [16] D. Milojkovic-Opsenica, P. Ristivojevic, TLC fingerprinting and pattern recognition methods in the assessment of authenticity of poplar-type propolis, J. Chromatogr. Sci. (2016) 1–7. [17] E. Guzelmeric, I. Vovk, E. Yesilada, Development and validation of an HPTLC method for apigenin 7-O-glucoside in chamomile flowers and its application for fingerprint discrimination of chamomile-like materials, J. Pharm. Biomed. Anal. 107 (2015) 108–118. [18] E. Reich, A. Schibli, High Performance Thin Layer Chromatography for the Analysis of Medicinal Plants, first ed., Thieme Stuttgart, New York, 2007. [19] Eigenvector Research, Inc Retrived March 15, 2015 from: http://www. eigenvector.com/software/pls toolbox.htm. [20] K.H. Wong, V. Razmovski-Naumovski, K.M. Li, G.Q. Li, K. Chan, Differentiating puerariae lobatae radix and puerariae thomsonii radix using HPTLC coupled with multivariate classification analyses, J. Pharm. Biomed. Anal. 95 (2014) 11–19. [21] N. Harbourne, J.C. Jacquier, D. O’Riordan, Optimisation of the extraction and processing conditions of chamomile (Matricaria chamomilla L.) for incorporation into a beverage, Food Chem. 115 (2009) 15–19. [22] F.N. Fonseca, M.F.M. Tavares, Validation of a capillary electrophoresis method for the quantitative determination of free and total apigenin in extracts of Chamomilla recutita, Phytochem. Anal. 15 (2004) 65–70. [23] G. Haghi, A. Hatami, A. Safaei, M. Mehran, Analysis of phenolic compounds in Matricaria chamomilla and its extracts by UPLC-UV, Res. Pharm. Sci. 9 (2014) 31–37. [24] M.R. Joharchi, M.S. Amiri, Taxanomic evaluation of misidentification of crude herbal drugs marketed in Iran, Avicenna J. Phytomed. 2 (2012) 105–112. [25] S. Sagi, B. Avula, Y.H. Wang, J. Zhao, I.A. Khan, Quantitative determination of seven chemical constituents and chemo-type differentiation of chamomiles using high-performance thin-layer chromatography, J. Sep. Sci. 37 (2014) 2797–2804. [26] B. Avula, Y.H. Wang, M. Wang, C. Avonto, J. Zhao, T.J. Smillie, D. Rua, I.A. Khan, Quantitative determination of phenolic compounds by UHPLC-UV-MS and use of partial least-square discriminant analysis to differentiate chemo-types of Chamomile/Chrysanthemum flower heads, J. Pharm. Biomed. Anal. 88 (2014) 278–288. [27] D. C¸alıs¸kan, Papatya Olarak Satılan Bitkiler Üzerinde Fitoterapötik C¸alıs¸malar. Master Thesis, Gazi University, Ankara, 2010. [28] L. Giansante, D. Di Vincenzo, G. Bianchi, Classification of monovarietal Italian olive oils by unsupervised (PCA) and supervised (LDA) chemometrics, J. Sci. Food Agric. 83 (2003) 905–911.